Does schema markup help AI search? What the evidence says

The short answer

Schema markup is not a switch for AI search. Nobody has shown that adding it gets a business mentioned more often in ChatGPT, Google AI Overviews, AI Mode or Perplexity. That does not make schema useless. It means you should judge it for what it does and not for what the sales pitch says it does.

This page sits inside our wider guide to AI search optimization. It covers one question: what schema does and does not do for AI features. For the markup itself, see schema markup for local business.

What Google says about schema and AI features

Google is the only major search company that has written this down in detail, so start there.

• No extra requirements. Google’s AI features page says there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary. A page has to be indexed and eligible to show in Search with a snippet.
• No special schema. The same page says there is no special schema.org structured data you need to add. Google’s guide to generative AI says the same: there is no special schema.org markup you need to add, and structured data is not required for generative AI search.
• No AI files. You do not need to create new machine-readable files, AI text files or markup to appear in these features. Google also says Google Search ignores llms.txt files. See llms.txt for the details.
• Match the visible text. Google advises making sure your structured data matches the visible text on the page.
• It is still SEO. Google’s guide says that optimizing for generative AI search is optimizing for the search experience, and thus still SEO.

One limit to keep in mind: these statements describe Google Search. They say nothing about how ChatGPT, Perplexity or Copilot work. That is why the independent tests below matter.

What the studies found

The Ahrefs test (May 2026)

Ahrefs, an SEO software company, published this study on May 11, 2026. It tracked 1,885 pages that added JSON-LD between August 2025 and March 2026. It compared them with 4,000 matched pages that never added it, using a difference-in-differences method over 30 days before and after. Its headline: adding schema produced no major uplift in citations on any platform.

Ahrefs lists its own limits:

• Every page in the study already had 100 or more AI Overview citations, so it cannot say whether schema helps pages AI has not seen yet.
• Pages that added JSON-LD often changed other things at the same time.
• It pooled all schema types, used a 30-day window and tested JSON-LD only.
• It called the small AI Overviews decline real but unexplained, so we would not read it as “schema hurts.”

Ahrefs also cites an experiment by searchVIU in which five AI systems, fetching pages in real time, extracted only the visible HTML and ignored JSON-LD and hidden markup. We have not opened that experiment ourselves, so treat it as Ahrefs’ report of it.

Ahrefs does not tell readers to stop using schema. It says that if you are already doing the rest of the work well, JSON-LD is not going to be the unlock, and that there are still many good reasons to use it.

The Search Atlas analysis

Search Atlas, which sells visibility tools, compared how much of a domain’s sampled pages carried schema with how visible the domain was in LLM answers from OpenAI, Gemini and Perplexity. It reports no link between schema coverage and citation frequency. Its caveats matter: it measured whether schema was present, not its type or quality; it shows correlation, not cause; and it tested three platforms. The page does not give the number of queries or domains, so we cannot judge the sample.

Reading them together

Both reports come from companies that sell SEO tools, and both are observational or quasi-experimental, not controlled lab trials. Still, they point the same way as Google’s statement. Be careful with the opposite claim, too. You will see articles saying that cited pages are more likely to have schema. That can be true without schema being the cause, because sites that bother with schema tend to be better built overall. Ahrefs’ matched test is designed to separate the two.

What Bing says

Microsoft’s February 2026 announcement of AI Performance in Bing Webmaster Tools gives publishers advice on getting cited in Copilot and Bing’s AI summaries. That post does not mention structured data. It lists clear headings, tables and FAQ sections, examples and sources, regular updates, consistent text, images and video, IndexNow, and Bing Places for local businesses. We found no Microsoft statement on schema in the pages we opened.

What schema still does well

Do not read the above as “delete your schema.” Google’s introduction to structured data says it uses structured data to understand the content of a page and to gather information about the web and the world in general. Adding it can also enable rich results, the enhanced listings in Search. In general, Google recommends JSON-LD.

Two cautions come with that:

1. Rich results are not guaranteed. Google says it does not guarantee that structured data will show up in search results.
2. Features come and go. Google ended FAQ rich results: they stopped appearing on May 7, 2026, and the documentation was removed on June 15, 2026. In 2025 it also dropped several other types, such as course info and special announcement. Check Google’s gallery before marking up something only for a rich result. Our guide to FAQ, review and breadcrumb schema covers what still works.

For a local service business, the practical value is simple. Schema gives your name, address, phone, hours, services and service area in a form software can read without guessing. It is a neat backup for the same facts that should already be in plain text on the page.

What to do instead of chasing schema

If your aim is to appear in AI answers, spend your first hours here. Schema comes after.

1. Put the facts in visible text. Services, area, price or “from” price, hours, how to book. AI tools and customers read the page text first.
2. Make sure the page is indexed and can show a snippet. That is Google’s only stated requirement.
3. Check that crawlers can reach you. See AI crawlers and robots.txt.
4. Keep your business details the same everywhere: your site, Google Business Profile and other listings.
5. Then add accurate schema for the facts already on the page, and test it with a validator. See how to test schema markup.

Keep markup honest. Google says not to mark up content that is not visible to readers of the page, and that structured data must be a true representation of the page content. A structured data manual action means a page loses eligibility for appearing as a rich result, though Google says it does not lose rank in web search.

How to test schema on your own site

Ahrefs suggests a simple design you can copy at small scale:

1. Pick 5 to 10 pages with some existing AI citations.
2. Pick 5 to 10 similar pages that will not get schema.
3. Record how often each group is cited today, across the AI tools you care about.
4. Add schema only to the first group. Change nothing else.
5. Check again after 30 days or more.

For most small businesses there is too little data for this to settle anything. That is a fair reason not to pay extra for it. For Google, Search Console’s Generative AI performance report shows how people discover your content through generative AI features. It does not tell you whether schema was the reason.

Questions to ask anyone selling “schema for AI”

• Which source says this schema affects AI answers? Google’s pages say it does not.
• Which types will you add, and which facts are already on the page that they describe?
• Will you validate it, and will you tell me when a feature changes?
• How will we measure a before and after?

Google also says third-party tools do not have access to its internal ranking or AI systems, and that using a service does not guarantee ranking success. Our guide on how to choose an AI SEO agency goes through more questions like these.

Nobody controls what ChatGPT, Google AI Overviews, AI Mode or Perplexity recommend. Schema is a way to state facts clearly, not a way to win placement.

How Magnet Mint handles schema

We are a Boston studio that builds websites for local service businesses and works with clients across the US, remotely. On the sites we build, we add schema for the facts a page already shows, validate it, and treat it as a tidy-up and not a promise. Our GEO and AEO audit reviews what schema you have against what your pages say. Price: from $750. Not included: domain and hosting, new photography, third-party software subscriptions (billed at cost), and AI usage above a plan allowance. To see how customers and AI assistants view your site today, ask for a free demo.

Published October 26, 2026. Updated October 26, 2026.